Thursday, September 23, 2004

Choosing Between Polls

There's been some chatter on the blog in regards to "choosing between polls," as if this was intrinsically problematic or ideologically driven. It's not. Here's my basic criterion for whether I believe a poll:

a) Is the source credible? That is, is the pollster not obviously in hoc to one or the other party, or someone who doesn't poll frequently enough for me to have confidence in the quality of their results;

b) Is the poll way off where the last presidential vote was?

c) Most important, is the poll not corroborated by other polls?

The Maryland poll, for example, is by a credible pollster, but criteria b and c are clearly violated. In the case of the latest Wisconsin poll, all three criteria are violated. So instead of casting aspersions, let's try to stick to the facts, and if you don't buy a poll, that's fine--explain why you agree with it or not, rather than falling back on partisan cheerleading. There are plenty of blogs for that sort of thing, and this ain't it.

2 comments:

Palooka said...

Zogby's state polling is consistently an outlier. From my casual obervance it's a good 6 or 8 points in favor of Kerry.

This criteria is far from objective. While the points are valid, they seem tailored to give you the result you want. One of your criteria is particularly glaring in this respect. By saying a poll cannot be way off of the last presidential race, you are saying you will reject anything which doesn't show a close race.

I think both Bush and Kerry supporters are likely to emphasize the good news, and ignore or marginalize the bad. It's a confirmation of the rightness of one's position. Stroking of the ego.

Klinker mentioned Ruy Texeira as someone who convincingly argued against the change in Party ID being plausible. But Mystery Pollster has found him in a stunning bit of hypocrisy. 3 months ago there was a LA times poll which showed a 7 point Kerry lead. The reason, of course, was a 13 point lead in Dem Party ID. He defended that as sound. Yet when recent polling shows a more plausible realignment, he scoffs at the possibility. Cocoon anyone?

Polling is an art as well as a science. So I guess I don't have any problem with interjecting subjective criteria like you have here. Your criteria, however, seems particularly suited to validating your preconceptions--that the race is close.

Anonymous said...

I've personally just been following the polls out of interest. I know not to take them too seriously but at the same time they can serve as a barometer of opinion.

As for which polls to trust, that's something that's way over my head. But I've been happy to follow the running poll averages that RealClearPolitics.com has been maintaining.